Perceptually lossless wavelet-based compression for very large oceanographic images

In this paper, we present a wavelet-based image compression scheme for perceptually lossless compression of oceanographic imagery. Wavelet coefficients are quantized within visual sensitivity of human eyes for different subbands to mask quantization distortions of compressed images. Large image files are divided into smaller blocks, each of which is compressed by an embedded tree-based coder which is capable of preserving smooth ocean textures, at the cost of complex land objects. In the mean time, a fast run-length filter detects and preserves image annotations at negligible computing time. Experimental results show that our coder produces perceptually identical image content on different display devices at a compression ratio 2-5 times higher than typical lossless compression schemes.